ClawHub
Back to skill

Security audit

Principle Comparator

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only text comparison skill, with one minor documentation inconsistency about whether review notes might be written to a file.

Reasonable to install for comparing text sources. Only provide content you are comfortable having processed by your configured agent model, and treat any `requires_review.md` output as ambiguous unless the publisher clarifies whether it is just a label or an actual file write.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The skill documentation makes conflicting claims about data handling: it says the skill does not write files to disk, but later states drift cases may be added to `requires_review.md`. This inconsistency can mislead operators about persistence behavior, which matters for sensitive user content and audit expectations even if no direct file-write primitive is shown in the skill itself.

VirusTotal

61/61 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.